Yeah that's what was happening with my friends car, the O2 would cool off too much during light load conditions. What appeared to be happening was; once the O2 started to cool off, the ECM didn't detect the O2 readings crossing the upper O2 boundary, and add to the fuel trim(increase BLMs) and essentially make the problem worse since adding all the extra fuel would only further cool the exhaust. IIRC, there's a qualifier that will measure the speed of the cross counts in either direction, and if the O2 values do not move fast enough in either direction, the ECM will give it a little nudge(add/subtract fuel trim) to get it moving faster. Anyway, eventually almost all the BLM cells were above 140 and the car would bog during an AE/PE event. Depending on how the TT chip is configured, the BLM may not reset to 128 during PE and if the BLMs are high that could attribute to an overly rich bogging condition.
Its ok if the O2 crosses the stoich boundary(.450V) at idle, in OL, as long as the engine likes it. Crosscounts are normal and have no effect in OL.
Was the engine using a Bosch O2 sensor? Those seem to be garbage sensors(IMO) and not operate properly more often than not. It also may be a good time to check for exhaust leaks, a large enough leak will skew the O2 readings at the sensor, significantly, and not head the sensor up enough to operate properly. If the Denso doesn't seem to remedy the issue, a 3 wire heated unit may be needed.
HTH
